When People Turn On the TV, Are They Actually Watching?
Advertisers pay hefty fees to run commercials during TV programs, but a study of baseball telecasts co-authored by Yale SOM’s Kosuke Uetake suggests that relatively few viewers actively look at the screen. The researchers also found that viewers perk up during suspenseful moments, suggesting that changes to advertising or even the structure of baseball games could make ads more effective.
An electronics store in Tokyo during the Olympics in August 2021.
The arrival of the smartphone has changed the way we watch TV. Instead of being glued to the big screen, you might watch a show for a few minutes, check your text messages, return to the TV, and then get distracted by a game on your phone.
These shifts in our viewing habits have raised concerns for advertisers who are shelling out big bucks to run commercials. “What are they really paying for?” asks Prof. Kosuke Uetake. In the United States, the fees for ads are typically based on ratings, but those data don’t account for people getting distracted. “If the rating is high, the TV is on, but nobody is watching,” then paying millions of dollars for a commercial “doesn’t really make sense,” he says.
Uetake and his colleagues took advantage of a new type of data in Japan that tracked how people’s attention waxed and waned as they watched TV. They focused on baseball games, which are hugely popular in that country, and determined when viewers were actually watching and when their attention drifted.
The team found that when people sat in front of the TV during a game, less than one-quarter of them were actively watching the program during any given moment. And they looked at the screen even less during commercials.
On the other hand, the researchers found that suspenseful moments increased people’s attention levels. “Suspense plays an important role,” says Uetake, who collaborated with Xiao Liu at New York University and Matthew Shum of Amazon on the research. That means that shortening baseball games could potentially boost attention to both games and commercials. When the action is compressed into less time, “each moment becomes more important,” he says.
Traditionally, data on TV viewership has been collected through audience measurement devices. Participating households attach the device to their TV, which automatically detects which program is playing. But people have to manually press a button to indicate if they’re watching or not, which introduces possible errors. For instance, someone might wander off to the kitchen five minutes after hitting the button and not actually watch the show.
A friend of Uetake’s helped create a company called TVision that collects data in a different way. With this technology, participants still attach a device to their TV, but it uses facial recognition to automatically detect if each household member was watching. The device also tracks eye movements to determine if they are paying attention or looking at something else.
Uetake’s team decided to study baseball game viewing because they ultimately wanted to analyze how different elements of a program, such as suspense and surprise, affected people’s attention levels. Capturing and scoring those elements during, say, a movie or sitcom would be tricky and subjective. But baseball games have a common structure and could be coded in a more uniform way.
The researchers started by analyzing data on 1,151 people in Japan who watched part of at least one baseball game in 2018. For each minute of the games and commercials, they determined if each person was “actively attentive,” meaning that their eyes were on the screen most of the time; only “passively attentive,” meaning they were in front of the TV but looking somewhere else most of the time; or not tuned into the game, meaning they were watching a different program or away from the TV.
For any given minute of a game, an average of about one-quarter of participants were in front of the TV with the game on. But among those viewers, only 22% of them were watching actively; the rest weren’t paying much attention.
During commercials, the level of passive attention didn’t change much, suggesting that most participants weren’t walking away or changing the channel. But one-quarter of the people who had been actively watching the game stopped focusing on the screen.
The team then wanted to know how moments of suspense or surprise changed people’s viewing behavior. To find out, they obtained information on 877 games from the company DataStadium. The time-stamped data set tracked what happened after every pitch, such as a home run or out.
Based on those details, the researchers gave each minute a score for suspense and surprise. For example, if all the bases were loaded, that moment rated high on suspense. If the losing team suddenly hit a home run, that play scored high on surprise. The team then determined what viewers did during and shortly after each of those moments.
Both passive and active attention went up when the game was suspenseful. But surprising moments had less of an effect on viewers’ attention levels afterward. Once the big moment has passed, people might think, “Oh, maybe we can go to the restroom, or maybe we can check email,” Uetake says.
The researchers also considered what might happen if the rules of the game were changed. For instance, Major League Baseball in the U.S. has floated the possibility of shortening games from nine to seven innings.
Based on a predictive model of baseball games, the team surmised that games would have more suspenseful and surprising moments if they were shorter. Essentially, some of the boring parts would be eliminated, so the exciting moments were concentrated into a shorter time. They estimated that active attention to both games and commercials would go up by about 2.6%. And passive attention to ads would increase too, by 2.0%.
Uetake speculates that during longer games, people might tend to skip commercials. Perhaps with shorter games, they might be more inclined to watch the whole game straight through, including the ads.
The findings also suggest that the structure of baseball broadcasts, in which most ads come between innings of play, might not lend itself to effective advertising. Companies might get more bang for their buck if their ads ran during suspenseful moments, Uetake says. “When is the optimal timing to introduce commercials?”